Last week cnn.com ran an article on an artificial intelligence (AI) designer and his attempts at a learning, household robot. He has had significant success, and he believes he is three years away from selling Zeno, a cartoonish looking robot that learns.
Zeno's programming is designed for him to learn, to read emotions, to talk, to walk, and be in all regards a robot pet.
Sentient AI will be one of the greatest challenges to human ethics and our understanding of life. We will be forced to face unanswerable and difficult questions: Can we truly create life? What is life? Will non-biologicals have equal worth as biologicals? Will it be OK for us to have robot "servants?" Will it be acceptable to shut-down sentient AI? I'm sure everyone will debate hotly: Do AI's have souls?
I do believe all sentient life is of value, and it is important for the future of the human race that we begin to discuss this issue. Too many times we have devalued life because it was different from us. In the 1800's, Native Americans and Blacks were thought by some to not have souls; rather, they believed that they were "almost" humans...worthy for slave labor and property. Color of skin, gender, and ethnicity have been sources for prejudice, but we will soon see another (and I believe very violent) prejudice against non-biologicals. It will be a fun discussion (to understate it a tad).
What do you think? Do you think we'll ever create sentient AI? Will you consider it life? Will its worth be that of your own?
7 comments:
I'm a big fan of technology and seeing it advance so you may find my skepticism odd.
With that said I don't think what you are talking about is life at all. Nothing man has put together truly has life it only mimics lifelike qualities. What really is the difference between that fancy robot and some computer programs? Yet we don't think of our PCs as having life, just that they are well programmed.
I also take issue with the term "self aware". I don't believe it is possible for a program to ever become truly self aware. We could program it to act like it is self aware, but to truly be like a child realizing that it can control its fingers, I don't see it. We have to program them to learn, react, emote, respond, but it will always just be a machine following orders.
Will it have a soul? No, how would it theologically speaking? Programming it to say "I" or "me" won't imbue it with one. I'm all for having them clean my house though, and would I feel sad if it fell down the stairs? Of course, as humans we assign human characteristics to inanimate objects all the time (stuffed animals and blankies to name a couple).
In an ironic twist I think the general populous will actually begin to regard them with more value than true human life. I mean it is so hard to get good help these days!
Good comments. I understand your reluctance to believe AI will truly become self-aware.
My understanding is that the brain is nothing more than a complicated biological computer.
I don't think WE will program life into the computer. I believe that we'll create a computer that will begin refining and reprogramming itself to a new level.
Here's a real thinker: Is that any less of a miracle?
I'll take a robot that does the laundry and the dishes. But the first time I catch it on the couch eating bon bons and watching TV it gets reprogramed! :)
Steph
I watched Bicentennial Man and for some reason the movie made me very blue. I think it was the rising and passing of generations that Robin William's character witnessed. I believe that would wear on all of us and perhaps aging and passing are blessings in disguise.
Never-the-less, if I remember correctly, doesn't the character's self awareness and depth of feeling reach its apex in the face of his own impending mortality?
It did me, too, as did AI. There's a tragedy in both stories that underlies this discussion.
Mark - you've hit on a lively debate with this topic! I hope we do see some interesting developments in robot technology in our lifetime. If you have a link to the CNN article I'd love to read it.
There are a number of guys here (Midway Games) who spend all day trying to program better AI. The one thing I know right now is if you want to have "smart" adaptive AI you have to pre-program the things it can "learn" to do. Basically you have to come up with all the options first and then tell it to evaluate the situation to pick the best option.
I will be amazed if I am ever shown an AI system that can develop solutions to problems the programmers could never have predicted. That's the day I'll stock up on EMPs... do you think they have them at Target?
I found that article here for those who are interested: CNN Article
Post a Comment